1 / 33

Caching and TLBs

Caching and TLBs. Sarah Diesburg Operating Systems COP 4610. Caching. Store copies of data at places that can be accessed more quickly than accessing the original Speed up access to frequently used data At a cost: Slows down the infrequently used data. Caching in Memory Hierarchy.

sandro
Download Presentation

Caching and TLBs

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Caching and TLBs Sarah Diesburg Operating Systems COP 4610

  2. Caching • Store copies of data at places that can be accessed more quickly than accessing the original • Speed up access to frequently used data • At a cost: Slows down the infrequently used data

  3. Caching in Memory Hierarchy • Provides the illusion of GB storage • With register access time

  4. Caching in Memory Hierarchy • Exploits two hardware characteristics • Smaller memory provides faster access times • Large memory provides cheaper storage per byte • Puts frequently accessed data in small, fast, and expensive memory • Assumption: non-random program access behaviors

  5. Locality in Access Patterns • Temporal locality: recently referenced locations are more likely to be referenced in the near future • e.g., files • Spatial locality: referenced locations tend to be clustered • e.g., listing all files under a directory

  6. Caching • Storing a small set of data in cache • Provides the following illusions • Large storage • Speed of small cache • Does not work well for programs with little localities • e.g., scanning the entire disk • Leaves behind cache content with no localities (cache pollution)

  7. Generic Issues in Caching • Effective metrics • Cache hit: a lookup is resolved by the content stored in cache • Cache miss: a lookup cannot be resolved by the content stored in cache • Effective access time: • P(hit)*(hit_cost) + P(miss)*(miss_cost)

  8. Effective Access Time • Cache hit rate: 99% • Cost: 2 clock cycles • Cache miss rate: 1% • Cost: 4 clock cycles • Effective access time: • 99%*2 + 1%*(2 + 4) = 1.98 + 0.06 = 2.04 (clock cycles)

  9. Implications • 10 MB of cache • Illusion of 4 GB of memory • Running at the speed of hardware cache

  10. Reasons for Cache Misses • Compulsory misses: data brought into the cache for the first time • e.g., booting • Capacity misses: caused by the limited size of a cache • A program may require a hash table that exceeds the cache capacity • Random access pattern • No caching policy can be effective

  11. Reasons for Cache Misses • Misses due to competing cache entries: a cache entry assigned to two pieces of data • When both active • Each will preempt the other • Policy misses: caused by cache replacement policy, which chooses which cache entry to replace when the cache is full

  12. Design Issues of Caching • How is a cache entry lookup performed? • Which cache entry should be replaced when the cache is full? • How to maintain consistency between the cache copy and the real data?

  13. Caching Applied to Address Translation • Process references the same page repeatedly • Translating each virtual address to physical address is wasteful • Translation lookaside buffer (TLB) • Track frequently used translations • Avoid translations in the common case

  14. In TLB TLB Translation table Data offsets into blocks (untranslated) Caching Applied to Address Translation Virtual addresses Physical addresses

  15. Example of the TLB Content

  16. TLB Lookups • Sequential search of the TLB table • Direct mapping: assigns each virtual page to a specific slot in the TLB • e.g., use upper bits of VPN to index TLB

  17. Direct Mapping if (TLB[UpperBits(vpn)].vpn == vpn) { return TLB[UpperBits(vpn)].ppn; } else { ppn = PageTable[vpn]; TLB[UpperBits(vpn)].control = INVALID; TLB[UpperBits(vpn)].vpn = vpn; TLB[UpperBits(vpn)].ppn = ppn; TLB[UpperBits(vpn)].control = VALID | RW return ppn; }

  18. Direct Mapping • When use only high order bits • Two pages may compete for the same TLB entry • May toss out needed TLB entries • When use only low order bits • TLB reference will be clustered • Failing to use full range of TLB entries • Common approach: combine both

  19. TLB Lookups • Sequential search of the TLB table • Direct mapping: assigns each virtual page to a specific slot in the TLB • e.g., use upper bits of VPN to index TLB • Set associativity: use N TLB banks to perform lookups in parallel

  20. VPN VPN PPN PPN hash VPN VPN PPN PPN VPN VPN PPN PPN = = Two-Way Associative Cache VPN

  21. VPN VPN PPN PPN hash VPN VPN PPN PPN VPN VPN PPN PPN = = Two-Way Associative Cache VPN

  22. VPN VPN PPN PPN hash VPN VPN PPN PPN VPN VPN PPN PPN = = Two-Way Associative Cache VPN If miss, translate and replace one of the entries

  23. TLB Lookups • Direct mapping: assigns each virtual page to a specific slot in the TLB • e.g., use upper bits of VPN to index TLB • Set associativity: use N TLB banks to perform lookups in parallel • Fully associative cache: allows looking up all TLB entries in parallel

  24. hash = = = Fully Associative Cache VPN VPN PPN VPN PPN VPN PPN

  25. hash = = = Fully Associative Cache VPN VPN PPN VPN PPN VPN PPN

  26. hash = = = Fully Associative Cache VPN VPN PPN VPN PPN VPN PPN If miss, translate and replace one of the entries

  27. TLB Lookups • Typically • TLBs are small and fully associative • Hardware caches use direct mapped or set-associative cache

  28. Replacement of TLB Entries • Direct mapping • Entry replaced whenever a VPN mismatches • Associative caches • Random replacement • LRU (least recently used) • MRU (most recently used) • Depending on reference patterns

  29. Replacement of TLB Entries • Hardware-level • TLB replacement is mostly random • Simple and fast • Software-level • Memory page replacements are more sophisticated • CPU cycles vs. cache hit rate

  30. Consistency Between TLB and Page Tables • Different processes have different page tables • TLB entries need to be invalidated on context switches • Alternatives: • Tag TLB entries with process IDs • Additional cost of hardware and comparisons per lookup

  31. Relationship Between TLB and HW Memory Caches • We can extend the principle of TLB • Virtually addressed cache: between the CPU and the translation tables • Physically addressed cache: between the translation tables and the main memory

  32. VA PA PA PA data data data data Virtually addressed cache Physically addressed cache PA PA VA PA data data data data TLB Translation tables Data offsets inside blocks (untranslated) Relationship Between TLB and HW Memory Caches

  33. Two Ways to Commit Data Changes • Write-through: immediately propagates update through various levels of caching • For critical data • Write-back: delays the propagation until the cached item is replaced • Goal: spread the cost of update propagation over multiple updates • Less costly

More Related